AUTHOR=Campi Cristina , Parkkonen Lauri T., Hari Riitta , Hyvärinen Aapo TITLE=Non-linear canonical correlation for joint analysis of MEG signals from two subjects JOURNAL=Frontiers in Neuroscience VOLUME=Volume 7 - 2013 YEAR=2013 URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2013.00107 DOI=10.3389/fnins.2013.00107 ISSN=1662-453X ABSTRACT=We consider the problem of analysing magnetoencephalography (MEG) data measured from two persons undergoing the same experiment, and we propose a method that searches for sources with maximally correlated energies. Our method is based on canonical correlation analysis (CCA), which provides linear transformations, one for each subject, such that the correlation between the transformed MEG signals is maximized. Here, we present a nonlinear version of CCA which measures the correlation of energies. Furthermore, we introduce a delay parameter in the model to analyse, e.g., leader-follower changes in experiments where the two subjects are engaged in social interaction.